Cognitive edge-cloud with serverless computing

ongoing

EDGELESS is set to efficiently operate serverless computing in extremely diverse computing environments from resource-constrained edge devices to highly-virtualised cloud platforms. By taking advantage of AI/ML solutions, it will enable automatic deployment and reconfiguration to fully exploit compute resources available on clusters of nearby edge nodes. EDGELESS will define novel orchestration systems that provide a flexible horizontally scalable compute solution able to fully use heterogeneous edge resources, while preserving vertical integration with the cloud and the benefits of serverless, including its application programming model. It will address edge systems at design stage, particularly targeting low-latency, high-reliability applications with computationally-intensive tasks, requiring specialised hardware or a trusted environment. This ambitious challenge will be met via distributed computing solutions to partition the edge environment in clusters, each managed as a local decentralised serverless platform. In each cluster, orchestration and scheduling of jobs will run smoothly thanks to real-time monitoring of short-term load/network/energy conditions and anticipatory AI-powered algorithms to manage lightweight virtualised lambda executors, e.g., unikernels. Environmental sustainability will be boosted by dynamically concentrating resources physically (e.g., by temporarily switching off far-edge devices) or logically (e.g., by dispatching tasks towards a specific set of nodes), at the expense of performance-tolerant applications. Clusters will cooperate with each other and with all the layers in the edge-cloud continuum to compose complex applications on-demand through a FaaS paradigm. EDGELESS innovations will be validated through testbeds (near-edge MEC and two small-device lab setups), integrated through a federated edge-cloud infrastructure, and three pilots: Autonomous Smart City Surveillance, Internet of Robotic Things, and HealthCare Assistants.